Performance of KEKB  with Crab Cavi5es 

Y. Funakoshi for KEKB accelerator Group 

KEK 

Outline

•  Overview of KEKB •  Crab cavity and crab crossing scheme 

•  Performance of KEKB with crab crossing 

•  Some experience of crab cavity opera5on with beams 

•  Summary and future plans

Overview of KEKB

Superconducting cavities (HER)

e-

e+

ARES copper cavities (LER)

8 GeV e- 3.5 GeV e+

Linac e+ target

ARES copper cavities (HER)

Belle detector

KEKB B-Factory

TRISTAN tunnel

• Circumference: • 3016m

• Beam energy • 3.5 GeV (e+; LER) • 8.0 GeV (e-; HER) • Ecm = 10.58GeV (Υ(4S))

• Beam Currents*

• 1.8A [1.62A] (2.6A) (LER) • 1.34A [0.95A] (1.1A) (HER)

• Number of Bunches: 1585/ring (~5000) • Horizontal crossing Angle:

• 22mrad or crab crossing • Peak Luminosity

• 1.0 x 1034cm-2s-1 on May 09 2003 • 1.71 x 1034cm-2s-1 (record w/o crab) • 1.61 x 1034cm-2s-1 (record w/ crab)

• Physics: • B physics (Asymmetric) (Belle)

• Integrated Luminosity: •  Total: >850 fb-1 •  ~ 1fb-1/day (record: 1.23 fb-1/day)

Crab cavi5es 1 for each ring

*Beam currents: [ ]: w/ crab, ( ):design

Peak luminosity

Daily integrated luminosity

Daily efficiency

Crab cavity and crab crossing scheme

Crab crossing scheme: some history •  The crab crossing scheme was proposed in 1988 by R. Palmer. 

–    Idea to recover the head‐on collision with the crossing angle for linear colliders.  

•  Also in ring colliders –  Oide and Yokoya showed that the synchro‐betatron coupling terms 

associated with the crossing angle are canceled by the crab crossing (1989). 

•  KEKB adopted a horizontal crossing angle of 22mrad. 

•  The crab cavi5es were considered as backup devices in KEKB design. 

•  The KEKB crab cavity was first designed by K. Akai at Cornel (1991‐1992) 

•  R&D of crab cavity started at KEKB in 1994. K. Hosoyama: Development of the KEK-B Superconducting Crab Cavity Thursday26 Morning Session in the Maestrale Auditorium

Horizontal crossing angle  KEKB

 PEP‐II

•  KEKB adopted 22mrad horizontal crossing angle. •  Easier beam separa5on 

•  Less SR background 

•  Less Luminosity‐dependent background 

•  Space for solenoid compensa5on •  Less parasi5c collision 

•  Possible demerit •  Synchro‐betatron resonance 

‐> We con5nued development of crab cavity as a backup device. 

•  Observa5on •  No serious effects of s‐b resonance 

•  We got ξy of 0.056 w/o crab cavity. 

Crab crossing at KEKB Head-on (crab)

First proposed by R. B. Palmer in 1988 for linear colliders.

  Crab Crossing can boost the beam-beam parameter higher than 0.15 ! (K. Ohmi)

22mrad crossing angle

Strong-strong beam-beam simulation

νx =.508

Head-on } ξy ~0.15

 Ader this simula5on appeared, the  development of crab cavi5es was revitalized. 

Structure of crab cavity Top View 

Input coupler Magnetic Shield　（ Jacket Type ）

80 K LN2 Radiation Shield

Coaxial Coupler

Stub Support

Bellows Liq. He Monitor Port

RF Absorber

Frequency Tuning by Adjusting Distance

Crab Mode Reject Filter

RF Absorber

I.D. 240 I.D.100

Liq. He

crab mode: TM110: By on beam axis lower mode: TM010: dumped through coaxial coupler

Finally two crab cavities were installed in KEKB,�one for each ring in February 2007!

HER (e-, 8 GeV) LER (e+, 3.5 GeV)

…..after 13 years’ R&D from 1994

Single crab cavity scheme •Beam tilts all around the ring. •z-dependent horizontal closed orbit. •tilt at the IP:

  1 crab cavity per ring.   saves the cost of the cavity and cryogenics.   avoids synchrotron radiation hitting the cavity.

Beam was indeed tilted! Observation with Streak Cameras (H. Ikeda et al)

inside of the rings

outside of the rings

long

itud

inal

horizontal

The streak camera

HER LER

Performance of KEKB with crab crossing 

        SuperKEKB (1.53mA2)  Ibunch (LER) = 1.87 mA  Ibunch (HER) = 0.82 mA 

Specific luminosity with crab crossing

Beam‐beam simula5on (strong‐strong)

crab crossing (beam study)

22 mrad crossing

crab crossing (physics run)

Machine parameters Nov. 2006 (w/o crab) Mar. 2008 (with crab)

Units LER HER LER HER

Circumference 3016 m

Hor.  emijance 18 24 15 24 nm

Beam current 1662 1340 1605 934 mA

# of bunches 1388+1 1584 + 1

RF frequency 508.88 MHz

RF Voltage 8.0 15.0 8.0 13.0 MV

νs ‐0.0246 ‐0.0226 ‐0.0240 ‐0.0204

νx / νy 45.505/43.534 44.509/41.565 45.505/43.567 44.509/41.596

βx* / βy* 59/0.65 56/0.59 90/0.59 90/0.59 cm

α (mom. compact.) 3.31 x 10‐4 3.38 x 10‐4 3.17 x 10‐4 3.38 x 10‐4

ξx / ξy 0.117/0.105 0.070/0.056 0.099/0.097 0.119/0.092

Beam life 110@1600 180@1340 94@1605 158@934 min.@mA

Luminosity 1.712 1.610 1034cm‐2s‐1

Beam-beam parameter

[mA] 

:experiments

ξy~0.093 (HER) (April 3 2007) 

Specific Luminosity and beam-beam parameter

Crab crossing • 49-sp βx*=80, 84cm εx=18, 24 nm • 3.5-sp βx*=80cm • 3.06-sp βx*=80cm • 3.06-sp βx*=90cm

22 mrad crossing

y=-16.35x+26.54 Green Ratio=100% 

Green line

ξy~0.093 (HER) (April 3 2007) 

Why the specific luminosity drops faster than expected?

•  Speculations:  Too many tuning parameters to find out an optimum set?  Short beam lifetime prevents us from approaching a

better parameter set?  Dynamic-β and dynamic emittance effects due to beam-beam?  Beam lifetime decrease dependent on horizontal orbit offset at IP

 Synchro-betatron resonance near 1/2 integer?  Some unknown fast noise?  Crosstalk between beam-beam and lattice non-linearity?  Vertical crab at IP?  and more….

Too many tuning parameters?

•  Many tuning knobs are tuned by scans only on the luminosity and the beam sizes and the beam lifetime.

•  Each scan takes a long time, typically ~30min. •  During beam operation, the operators are almost always doing some scan. •  Is it possible to reach an optimum set of parameters with this method? •  We still suspect this possibility.

Dynamic-β and dynamic emittance �by beam-beam (calculation)

LER

The focusing force of the beam-beam interaction not only squeezes the beam at the interaction point, but increases the emittance drastically.

with/without beam-beam effects

LER

HER

Deformation of β- function all around the ring due to beam-beam effect (“dynamic beta”)

QC2L 

QC2R 

QW4NP.1 

QW4NP.2 

QW4OP.2 

QW4OP.1 

MD06H1 

MD06H2 

MD06H3 

MD06H4 

MD03H1 

MD03H2 

MD03H3 

MD03H4 

Beam size calculation with dynamic beam-beam effects

@crab (aperture < 5 σx)

βx* = 0.8m

βx* = 1.5m κ = 1%

βx* = 1.5m κ = 1.3%

w/o crab βx* = 0.8m 

Machine study βx* = 1.5m 

Trial of higher βx*

βx*: 0.9 ‐> 1.5m   βx@crab: 199 ‐> 109m (LER)  βx@crab: 160 ‐> 97m   (HER)

•  The (HER) beam current seems to be limited by the short life time of the LER beam.

•  The (LER) beam lifetime is very asymmetric with respect to H offset.

Collision center given by the beam-beam kick

LER lifetime

Mysterious lifetime asymmetry with respect to sign of hor. offset!

Horizontal offset at IP and crossing angle

Beam life

Beam size

Luminosity

Beam‐beam simula5on Horizontal offset scan: experiment with rela5vely small beam current

horizontal offset

Crab Vc scan (experiment in physics run)   Luminosity boost by crab crossing disappears�with 2 mrad crossing angle.   Luminosity boost by crab crossing disappears�with ~40 µm horizontal offset.   Typical value of horizontal offset in physics �experiment is ~15 µm, which is obtained by offset�scan.   This kind of horizontal offset depending on beam �current can degrade the specific luminosity.   Some luminosity boost by the crab crossing is�actually observed by crab Vc scan.

Synchro-betatron resonance •  The horizontal tune is set nearby

the half integer resonance and its synchrotron side bands.

•  On the resonances, some harmful effects are observed.

–  Single-beam beam size blowup –  Tow-beam beam size blowup –  Beam lifetime reduction (or beam loss)

•  The resonance is stronger in HER where no local chromaticity correction is installed.

•  Strength of the resonance is strongly dependent on a choice of sextupole setting. The luminosity also changes by changing the sextupole setting.

•  Even in the off resonance tunes, it affects the luminosity due to the tune footprint by the beam-beam?

•  We still suspect this resonance.

2νx + νs = integer

2νx + 2νs = integer

LER HER

Negative-α Optics •  Mo5va5on 

–  To weaken the synchro‐betatron resonance par5cularly in HER 

–  To shorten the bunch length •  Results 

–  We have succeeded to weaken the synchro‐betatron resonance line in HER. We could operate the machine with νx below the resonance lines. 

–  We have successfully shorten the bunch length of both beam. •  ~6mm ‐> ~4.5mm 

–  However, we found unexpectedly large synchrotron oscilla5on in LER due to the microwave instability and gave up the trial of the nega5ve‐α op5cs. 

–  Recently, we succeeded in suppressing the instability by increasing  the absolute value of α.  

2νx + νs = integer 2νx + 2νs = integer

2νx - νs = integer 2νx - 2νs = integer

νx: .5112, .5224 with given νs ~ -.0224

LER Optics (cell)

Negative α Positive α

H. Koiso

LER Optics

Negative α Positive α (now)

2005/June/06

Noise of feedback system affected luminosity? (22mrad crossing angle)

 We once found that the luminosity increased by lowering the gain of bunch-by-bunch feedback system (LER vertical) in case of 22mrad crossing angle.  It seemed that some noise from the FB system affected the luminosity.  At present, no remarkable effect of FB gain is observed.

original gain

Latest beam-beam simulation

0

5

10

15

20

25

-0.5 0 0.5 1 1.5 2 2.5 3

!""#$#%&#'("")'*+(,#,(#'"-.!""#$#%&#'("")'*+(,#'"-.!""#$#%&#'("")'*+(,/#'"-./#0#'("")'*+(,!""#$#%&#'("")'*+(,/#1('-2#'"-./#0#'("")'*+(,3*"(,456*"(,4#'"-.3*"(,456*"(,4#778"-9#!:;)"+8),*

Spec

ific

lum

inosi

ty /

bunch

[10

30cm

-2s-1

]

I+ x I

- [mA2]

K. Ohmi

Preliminary!!

Some experience of crab cavity opera5on with beams 

Phase stability of crab cavities •  Two measurements with different signals (cavity pick up signal and

signal from phase detector of PLL) give a consistent result. •  Phase fluctuation faster than 1 kHz is less than ±0.01°, and slow

fluctuation from ten to several hundreds Hz is about ±0.1°. •  They are much less than the allowed phase error obtained from the

beam-beam simulations for the crabbing beams in KEKB.

Span 200 kHz Sideband peaks at 32kHz and 64kHz.

Span 10 kHz Span 500 Hz Sideband peaks

at 32, 37, 46, 50, 100 Hz. Phase detector signal. Beam current was 385mA (HER) and 600 mA (LER).

According to beam-beam simulation by K. Ohmi, allowed phase error for N-turn correlation is 0.1×√N (degree). The measured phase errors are much smaller than the allowed values given by beam-beam simulation.

Spectrum around the crabbing mode measured at a pick up port of the LER crab cavity. Beam current was between 450 and 600 mA.

K. Akai

Beam opera5on with crab cavi5es

HER current

Luminosity

LER current

Crab detuned

Trip rate of crab cavities

Summary •  20 years ader they were ini5ally proposed, in February 2007 crab cavi5es 

are for the first 5me installed in an opera5ng collider, KEKB.  

•  The crab cavi5es at KEKB have been working much more stably than the ini5al expecta5on. 

–  They are presently being used in usual physics run (high beam current!!). 

•  The success of the development of the crab cavi5es is important, since they can be applied to other machines such as SR facili5es or an upgrade of LHC. 

•  With crab crossing, the ver5cal beam‐beam parameter of 0.093 was obtained. This indicates superiority of crab crossing scheme. 

•  However, the crab cavity at KEKB has not yet fully realized its poten5al capability in the sense that the specific luminosity is much lower than the beam‐beam simula5on at the high bunch currents. 

•  Finding the cause of this problem is very important for KEKB, since the design of SuperKEKB already counts the luminosity gain by the crab cavi5es.

Future plans •  We will con5nue the inves5ga5on on the low specific luminosity at high bunch currents. 

•  More beam‐beam simula5on (Ohmi) 

•  In the autumn run this year, the e+ and e‐ simultaneous injec5on may be realized at KEKB. It is expected that the beam opera5on with shorter beam life5me will be possible. Some luminosity gain is expected with this.   

•  KEKB maybe con5nue its opera5on also in the next fiscal year (Apr. 2009~). 

Spare slides

Downhill Simplex Method

Reflect

Expand

Contract+ Contract-

Shrink

1

2 3

Method of Minimization •  {1, 2, 3} 1(best)<2(next-to-the worst)<3(worst) •  Evaluate 3R •  If 3R<1,

•  If 3E<3R, {1, 2, 3E} : Expand , if not, {1, 2, 3R} : Reflect •  If 1<3R<2, {1, 2, 3R} : Reflect •  If 2<3R<3, Reflect 2 proposed by A. Hutton

•  If 3C+<3R, {1, 2, 3C+} : Contract+ , if not, {1, 2, 3R} : Reflect •  If 3<3R, Reflect 2

•  If 3C-<3, {1, 2, 3C-} : Contract- , if not, {1, 2S, 3S} : Shrink/Reflect2

H. Koiso 

5

10

15

20

25

0 0.5 1 1.5 2

L/I

+I - 1

0-3

0

I+I- (mA2)

HER 24nm 1%

Err-1, IP cor.Crabbing

Err-2, IP cor.Crabbing

No err Crabbing

βx* = 0.9m  

βx* = 1.5m βx* = 0.9m  LER Nikko new op5cs  βx* = 0.9m 

Machine study

42 

LER Nikko  New Op5cs 

Before 2/21(maintenance) βx max 199 m (νx,νy)=(45.505, 43.59)

After 2/21(maintenance) βx max 91 m (νx,νy)=(44.505, 43.59) Large β distortion in wiggler section

crab crab βx @ crab ~85 m as is before

βx βy

H. Koiso

Bunch length measurement

α= -2.55E-4 α= -2.55E-4 α = +3.31E-4 α  = -3.41E-4 α = +3.41E-4

H. Ikeda

Synchro‐betatron resonance in HER 

•  Posi5ve‐α – We could NOT operate under the resonance (2νx+νs=integer)  

• Negative-α –  We could operate

under the resonance (2νx- νs=integer)

The luminosity depends on the sextupole sewng 

Chromaticity tuning

- We often observed the horizontal size changed dependent on sextuple setting. - The luminosity is also improved with the chromaticity tuning.

LER

LER

νx νx

σx (µ

m)

σx (µ

m)

Calcula5on of beam‐beam parameter 

•  Reduc5on factor for beam‐beam parameter 

–  2 sources of reduc5on •  hourglass effect and finite crossing angle 

€

ξy = Rξ yξy0 ξy0 =re2πγ

βy*N

σ y* σ x

* +σ y*( )

€

Rξ y = 1+z /2βy*

2

−∞

∞

∫ fy x,σ x,σ y( )ρ z( )dz

€

fy x,σ x,σ y( ) =kk −1

1− e−x 2

2σ x2 1k

+i π x

σ x 2 1− k 2( )w xσ x 2 1− k 2( )

− e

−x 2

2σ x2

w kxσ x 2 1− k 2( )

€

k =σ y

σ x

ρ z( ) =12πσ z

e−z 2

2σ z2

Montague’s factor

Calcula5on of beam‐beam parameter [cont’d] •  Reduc5on factor for luminosity 

–  Luminosity 

–  We use calculated values for σx* and calculate 

 σy* and ξy0 from observed luminosity. 

€

RL ≡LL0

=2πaebK0 b( )

a =βy*

2σ z

, b = a2 1+σ z

σ x* tanφ

2

€

L =14π

N +N−

σ x*σ y

* fcolRL